What is Microsoft Fabric?

TL;DR

Microsoft Fabric is an all-in-one analytics platform (SaaS) that combines data engineering, data warehousing, data science, real-time analytics, and Power BI — all sharing a single storage layer called OneLake. Think of it as Microsoft’s answer to Databricks, but with Power BI baked in and everything managed by one capacity you pay for.

The Big Picture

Microsoft Fabric architecture: data sources flow into OneLake, which feeds 7 experiences — Data Engineering, Data Warehouse, Data Science, Data Factory, Real-Time Intelligence, Power BI, and Data Activator

Before Fabric, Microsoft’s analytics story was scattered across Azure Data Factory, Azure Synapse, Azure Data Lake, Power BI, and more — each with its own billing, security, and storage. Fabric brings them all under one roof with one storage format (Delta Parquet in OneLake), one security model, one billing (capacity units), and one governance layer.

Key insight: Fabric is SaaS, not IaaS. You don’t manage VMs, clusters, or storage accounts. You buy capacity (CUs) and Microsoft handles the infrastructure.
Explain Like I'm 12

Imagine your school needs different things: a library for books (data storage), a science lab for experiments (data science), a newspaper team for reporting (Power BI), and a mail room for delivering packages (data pipelines). Before Fabric, each of these was a separate building with separate keys and separate budgets. Microsoft Fabric is like combining all of them into one mega-building where everyone shares the same hallways, the same security badge, and the same budget — and all the books are on the same shelf so nobody has to carry copies around.

Who is Microsoft Fabric For?

RoleWhat You Do in Fabric
Data EngineerBuild pipelines, create lakehouses, transform data with Spark notebooks or Dataflows Gen2
Data AnalystQuery the warehouse with T-SQL, build Power BI reports directly on lakehouse/warehouse data
Data ScientistTrain ML models in Spark notebooks, use MLflow for experiment tracking
Analytics EngineerBuild semantic models (formerly datasets), manage data quality, set up medallion architecture
BI DeveloperCreate Power BI reports and dashboards with Direct Lake mode for blazing-fast visuals

What You’ll Learn

Start Learning: Core Concepts →

Key Capabilities

CapabilityWhat It Means
OneLakeSingle storage layer for all data — like OneDrive for data. One copy, accessible by all engines.
LakehouseCombine files and Delta tables in one place. Query with Spark or SQL.
WarehouseFully managed T-SQL data warehouse with cross-database queries.
Data PipelinesADF-like orchestration — copy, transform, schedule, and monitor data flows.
Dataflows Gen2Power Query Online for no-code/low-code ETL that lands data into lakehouses.
NotebooksApache Spark notebooks supporting PySpark, Spark SQL, Scala, and R.
Real-Time IntelligenceIngest and query streaming data with KQL (Kusto Query Language) and Eventstreams.
Power BI (built-in)Create reports and dashboards directly on Fabric data with Direct Lake mode.
CopilotAI assistant across all experiences — generate code, DAX, SQL, and summaries.

Microsoft Fabric vs. Alternatives

FeatureMicrosoft FabricDatabricksSnowflake
ArchitectureSaaS lakehouse (OneLake)Lakehouse (Delta Lake)Cloud data warehouse
StorageOneLake (Delta Parquet)Delta Lake (open Parquet)Proprietary + Iceberg
BI toolPower BI (built-in)Partner connect (Tableau, etc.)Snowsight + partners
ETLPipelines + Dataflows Gen2Workflows + SparkSnowpipe + Streams/Tasks
ML / Data ScienceSpark notebooks + MLflowMLflow + AutoMLSnowpark ML
Real-timeEventstreams + KQLStructured StreamingSnowpipe Streaming
PricingCapacity Units (CUs)DBU-basedCredit-based
GovernancePurview integrationUnity CatalogHorizon (governance)

Test Yourself

Q: What problem does Microsoft Fabric solve that Azure Synapse didn’t?

Fabric unifies data engineering, warehousing, data science, real-time analytics, and BI into a single SaaS product with one storage layer (OneLake), one security model, and one billing model. Azure Synapse was one piece of a fragmented puzzle — you still needed separate services for Power BI, ADF, ADLS, etc.

Q: What is OneLake, and why does it matter?

OneLake is Fabric’s single, unified storage layer — think “OneDrive for data.” All Fabric workloads read and write to the same OneLake storage. This eliminates data duplication, simplifies governance, and means every engine (Spark, T-SQL, Power BI) can access the same data without copying it.

Q: How is Fabric different from Databricks?

Both are lakehouse platforms, but Fabric is a fully managed SaaS (no cluster management, single billing via CUs, Power BI built-in). Databricks gives you more control over Spark clusters and is cloud-agnostic (AWS/Azure/GCP). Fabric is Azure-only and tightly integrated with the Microsoft ecosystem (Power BI, Purview, Teams, M365).

Q: What are the main “experiences” inside Fabric?

Fabric has 7 main experiences: Data Engineering (Spark, lakehouses), Data Warehouse (T-SQL warehouse), Data Science (ML notebooks), Data Factory (pipelines + Dataflows Gen2), Real-Time Intelligence (Eventstreams + KQL), Power BI (reports + dashboards), and Data Activator (trigger actions on data conditions).

Q: What storage format does Fabric use internally?

Delta Parquet. All data in OneLake is stored as Delta tables (Parquet files + a transaction log). This gives you ACID transactions, time travel, and schema enforcement — the same open format used by Databricks Delta Lake.